Lj4040

您好，我在下载您的RoBERTa-zh-Large: [Google Drive ](https://drive.google.com/open?id=1yK_P8VhWZtdgzaG0gJ3zUGOKWODitKXZ)或 [百度网盘](https://pan.baidu.com/s/1MRDuVqUROMdSKr6HD9x1mw) ，PyTorch版本，Bert的PyTorch版直接加载， 我在下载这个文件时，下载到一半报错，报没有权限我将如何去获得权限，另一个问题是我看您下面写着可以直接加载，我是使用hfl/RoBERTa-zh-Large 直接加载还是需要如何去加载，希望您可以回复下，同时也谢谢您的回复

stage 2 training data problem

3

Phase 2 requires data such as Fce and W&I+LOCNESS, which is M2 file format after I have downloaded it. How can I convert this format into two parallel files? I...

data preprocessing

7

What do SOURCE and TARGER stand for in data preprocessing? Could you explain them? Thank you for your reply

labels = torch.arange(batch_size, device=input_ids.device) cont_loss = cont_crit(logits, labels) new_cont_loss = cont_crit(new_logits, labels) Hello! Look at your work and feel very good! So I tried to learn, in the process of...

I'm really interested in your work. @pemywei Could you share the source code with me? Very appreciate it. My email is:[[email protected]] Thank you very much

丢失文件

1

您好！看了您的论文觉得非常帮！我在下载完代码进行复现时，发现代码好像并不完整，比如loss那缺了好几个from simpletransformers.losses.focal_loss import FocalLoss from simpletransformers.losses.dice_loss import DiceLoss from simpletransformers.losses.tversky_loss import TverskyLoss 还有您安装包的版本可以分享一下吗？感谢您的回复